outside the box

Blog facelift in progress

2007-03-25T20:39:00.000-05:00

I've decided to do a blog facelift that I have been wanting to do for a while now.

First is the change in the blog template. I liked the look of the old template (Scribe) but the narrow post body made it problematic to format posts that include source code or screen output. The new template is a slightly modified version of the Stretch Denim Light template. As a result of the new template, I will be modifying some posts that include source code or screen output to make them look better with this new blog template. These modifications will be purely cosmetic in nature and subscribers can safely ignore them in their feedreaders if they have read the post before.

Next, I have finally started a blogroll (shame on me) :).

Finally, I added links of websites that I hope readers of this blog will find useful. If you are running Oracle on Solaris, you should definitely check the links to the Solaris Internals website.

OraSRP v3.0b - screenshots for discussion

2007-03-08T18:44:00.000-05:00

Update:
March 9, 2007
1. I changed the title of this post from OraSRP v3.0b - short program to approximate session elapsed time to OraSRP v3.0b - screenshots for discussion.
2. Added another screenshot of a short program to look for the minimum and maximum values of tim.

This entry is created to support some points being made in the comment section of the entry titled OraSRP v3.0b - questionable session elapsed time. It was created for the sole purpose of displaying ~~this screenshot~~ the screenshots below because it looks like blogspot does not support screenshots in the comment section.

1. A short program to approximate session elapsed time of the trace file being discussed.

2. A short program to find the minimum and maximum tim values of the trace file being discussed.

Comments are disabled for this entry. If you want to join the discussion, please go here.

OraSRP v3.0b - questionable session elapsed time

2007-03-03T14:05:00.000-05:00

Note to subscribers: If your reader does not properly display some formatted content like bold and italics, I suggest you visit the blog site itself.

For those who may just be passing by, this is a continuation of two previous blog entries:

If you have not read these two entries, I suggest that you do so first, specially the first one, since that is where the possible problem with the calculation of session elapsed time was initially described. I also suggest that you read Egor's comment for that entry.

As the title says, this entry is about an observation that I made regarding the discrepancy between the session elapsed time reported by OraSRP, and the contents of the raw Oracle Extended SQL Trace File (simply referred to as trace file from hereon). Earlier, I said that OraSRP reports a number that is around 16 seconds more than what I can readily see and approximate from the trace file. In his comment, Egor explained:

It depends on how to calculate elapsed time.
For example, OraSRP ignores 'tim' in 'PARSING IN CURSOR' lines.
Also, if trace file starts from WAIT lines then OraSRP will sum up elapsed times of such waits and substract it from the first 'tim' value.

Thank you, Egor, for that explanation. But please allow me to be more specific.

Session Start Timestamp
Here's a screenshot that shows the top of the trace file (sensitive information has been crossed out, please click to enlarge):

From that screenshot, I make the following observations:

There are no WAIT lines before the first 'PARSING IN CURSOR' line so the mention of WAIT lines in Egor's comments is not applicable to this specific situation.

Let's follow Egor's comment and ignore the tim value from the 'PARSING IN CURSOR' line (line 22). The next tim value after line 22 comes from the EXEC line (line 25). The value of tim for that line is 3928564402472. Those who are closely following the values of the timestamps may notice that the value of tim from the EXEC line (line 25) is less than the value of tim from the 'PARSING IN CURSOR' line. In this specific case, the EXEC tim is 488 microseconds (or 488/1,000,000 of a second) before the tim from 'PARSING IN CURSOR'. This may seem a little bit odd but if you happen to have the Optimizing Oracle Performance book, the reason why this is so is explained on pages 100-101 in the section titled "Oracle Release 9".

At this point, I think it is safe to make the following statements regarding the start timestamp:

The actual start timestamp can not be greater than 3928564402472. By that I mean that it can be less, but not 16,000,000 microseconds less.
I believe that the earliest timestamp that one can get from this specific tracefile is the one in the 'SESSION ID' line (line 19). In fact, the Optimizing Oracle Performance book describes this line as the line that "identifies the time and the session at which the first trace line was emitted" (page 104). Purists will probably disagree with me but personally, I can tolerate not using the value in line 19 as start time, specially in cases where the session elapsed time is big (like in this case where it is more than 14,171 seconds). I say this for the following reasons:
- The 'SESSION ID' time is in a different format and converting it to the tim format is not always straightforward. Of course, if one can figure out a reliable way to convert the 'SESSION ID' time to tim format, then this is the time that should be used as session start time.
- The difference between the times that are represented between lines 19 and 25 is likely very small, just like the difference between lines 22 and 25 is also small (488 microseconds). I can personally tolerate this difference knowing that it is a small one-time occurence. And in the context of what is being discussed here now, I can not imagine a valid case where this difference would go as high as 16 seconds (or 16,000,000 microseconds).

Session End Timestamp
Here's a screenshot that shows the bottom of the trace file (sensitive information has been crossed out, please click to enlarge):

From that screenshot, I make the following observations:

The last line that has a tim is the EXEC line (line 2156805) and the value of tim from this line is 3942736181428.
I said in the first post that the actual end timestamp can not be less than the tim value in the EXEC line above. By that, I implied that it can be more, but not 16,000,000 microseconds more. In this specific case, there are several WAIT lines that follow this EXEC line. The ela values for these WAIT lines must be accounted for.

The total amount of time that can be attributed to these WAIT lines is the sum of their ela values:


Line #    ela
------- -----
2156806     1
2156807   560
2156809   917
2156810     2
2156811   518
------- -----
Total   1,998

jforonda@u20$ bc
1+560+917+2+518
1998
jforonda@u20$

To get the session's end timestamp, simply add the last tim value (3942736181428 from line 2156805) to the total ela of the WAIT lines above (see Clock Walk Formulas, p101-102, Optimizing Oracle Performance).

jforonda@u20$ bc
3942736181428+1998
3942736183426
jforonda@u20$

To the best of my knowledge, that's the maximum value of the end timestamp that one can get from this specific tracefile. If anyone can see anything that I've missed, I will welcome your corrections.

Session Elapsed Time

The session elapsed time should just be the difference between the end timestamp and start timestamp:

jforonda@u20$ bc
scale=6
(3942736183426-3928564402472)/1000000
14171.780954
jforonda@u20$

So there you go. That is how I would calculate this specific trace file's session elapsed time. Is this how others would do it? If what I presented above is wrong, where did it go wrong? If the method above is wrong, is it so way off to a point that it can not be relied on? Is it even reasonably close enough to be within the bounds of what is acceptable? Comments are welcome, whether you agree or disagree. Moderation is on to control spam.

Assuming that the calculation above is correct (or at least reasonably close enough to what it should be), it means that nothing much has changed from the initial observations that I made in the first post -- that is, there is a discrepancy of around 16 seconds from the value reported by OraSRP:

jforonda@u20$ bc
scale=6
14187.9512-14171.780954
16.170246
jforonda@u20$

Again, assuming that the calculation above is correct, some people might then ask: why bother with 16 seconds specially since it is just around 0.11% of the total? I'm very interested to hear what others think about this. Does this 16-second discrepancy tell you anything? At the very least, does it suggest anything? What kinds of questions come up on your mind as a result of this? If you are a person who does Oracle performance analysis and improvement in high profile situations, is this 16-second discrepancy acceptable to you? If you have some level of understanding of the trace file and you have a programming background at the same time, is this discrepancy acceptable to you? Why? Why not? Again, your comments are welcome. If there are no comments, I will take this up in a future post when my schedule permits.

If you find this series of posts was worth your time, I would be honored if you subscribe.

OraSRP author replies

2007-03-01T14:42:00.000-05:00

Note to subscribers: If your reader does not properly display some formatted content like bold and italics (I know that at the very least, Bloglines has this problem), I suggest you visit the blog site itself.

First of all, I would like to thank OraSRP's author, Egor Starostin, for dropping by and making a comment on yesterday's blog entry, OraSRP v3.0b - a very short initial impression. I've decided to put my reply in a separate blog entry so that readers who do not subscribe to comment feeds can pick it up.

With regards to my observation that the number of lines that OraSRP reports is not equal to the actual number of lines in the raw trace file, Egor says:

OraSRP ignores some lines like
'===' or '*** 2007-02-28 12:45:08.424' and doesn't report them as 'processed'. So, 'Total lines processed in trace file' != 'Total lines in trace file'.

That's good to know, and I suspected that much. But without that clarification, a different line count may mean that there are some relevant lines that are being missed. And when relevant lines are missed, it then becomes possible that the output may be affected to a point that it no longer presents a correct profile. When the profile is not correct, then that may lead people to try to solve problems that are not the real ones to begin with. This is very important to me because the real strength of a correct profile, at least as far as I'm concerned, is its ability to pinpoint the root cause of a problem.

Of course, I have no problems bypassing irrelevant lines and since this is already being done, I respectfully suggest that the number of lines in the raw trace file still be reported, even as a separate line item, so that people don't get thrown off by differing numbers.

Now, let's double check the numbers and take into account the types of lines that Egor mentioned above. Let me deal with the '^===' lines first:

jforonda@u20$ egrep '^===' 4071.trc | head -5
=====================
=====================
=====================
=====================
=====================
jforonda@u20$

That looks good enough. Now, let's count them:

jforonda@u20$ egrep '^===' 4071.trc | wc -l
2165
jforonda@u20$

So we have 2,165 lines that fit this criteria.

For the date/timestamp lines:

jforonda@u20$ egrep '\*\*\* 200' 4071.trc | head -5
*** 2004-10-18 19:51:28.453
*** 2004-10-18 19:55:50.302
*** 2004-10-18 19:56:07.899
*** 2004-10-18 19:56:48.561
*** 2004-10-18 19:57:41.135
jforonda@u20$

jforonda@u20$ egrep '\*\*\* 200' 4071.trc | wc -l
106
jforonda@u20$

That's another 106 lines. If we add these two numbers to the total reported by OraSRP, we get a total of 2,152,611:

jforonda@u20$ bc
2150340+106+2165
2152611
jforonda@u20$

This means that there are still 4,239 lines that are not accounted for:

jforonda@u20$ bc
2156850-2152611
4239
jforonda@u20$

This is still a lot of lines that are not unaccounted for. Now, Egor's comment said "OraSRP ignores some lines like..." and that makes me think that there are other types of lines, aside from those that were already mentioned, that are being bypassed on purpose. If Egor gives me a complete list of the types of lines that are being bypassed on purpose, I'd be happy to check my trace file.

With regards to the different totals between the “with idle” and "without idle” runs, Egor said:

As for different times with/without idle waits recognition -- I believe it's a bug. Unfortunatey, I can't reproduce it yet, so I can't fix it too.

In as much as I want to help, I will say now that I can not give this trace file to Egor. As most people reading this probably know, 10046 trace files contain confidential information. This is also the reason why I crossed out the instance and node names from the screenshots that I posted yesterday. If there are readers who can share their trace files to Egor, I suggest that they try to do what I've done here and see if they can reproduce the problems that I have described.

Now, there's still that issue about the total elapsed time but I have to attend to other things so I will stop for now. I will post another entry to talk about the elapsed time within the next day or so. If you want to follow this topic, it will make it easier if you subscribe. :)

OraSRP v3.0b - a very short initial impression

2007-02-27T23:31:00.000-05:00

Update: This has evolved into a series of posts so I added links at the bottom to guide readers who may want to read the entire series of posts.

Last Friday, February 23, Doug Burns mentioned in his blog that OraSRP v3.0b (beta) has been released. I have a special interest in the Oracle Extended SQL Trace File because I have used it to help me solve a lot of performance problems. My interest with the trace file started after reading the excellent book Optimizing Oracle Performance by Cary Millsap and Jeff Holt of Hotsos.

Anyway, I've tried earlier versions of OraSRP but I had some problems. Off the top of my head:

It could not finish processing certain files. This has nothing to do with speed. It had something to do with OraSRP not being able to finish processing large files with a lot of BIND lines in them. I had two big files (44+ and 66+ million lines) where OraSRP would just stall. I ran OraSRP against these files on several platforms (Solaris, Linux, Windows) and I had to kill them after two days. The workaround to this type of problem was to bypass or limit the processing of BIND lines.

I compared the output to that of the Hotsos profiler (v4.0.16). There were some numbers that were too far apart, so far that they can not be attributed to rounding off. This does not mean, of course that OraSRP was wrong. I'm just saying the outputs of the two profilers were vastly different in some sections.

Again, those two bullet points are for an earlier version of OraSRP. I also want to make clear that I did not have time to prove which output is correct and so I could not conclude that OraSRP was wrong; after all, it is entirely possible that there are some bugs in the Hotsos profiler itself. But I have to say that the big differences in the numbers made me very uncomfortable not only because I know that the people at Hotsos wrote THE book on the Oracle Extended SQL Trace File, but also because I have used earlier versions of the Hotsos profiler to help me solve some very nasty performance problems at some of Sun's big customers.

And so it is natural for me to want to see what is new with OraSRP v3.0b. This new version is a complete rewrite from Python to the D programming language (not to be confused with a similarly named language for Solaris DTrace). It is also in binary-only format for Linux and Windows.

These days, my main workstation runs Solaris and I no longer have a machine that runs Linux so I had to use the Windows version. Here's a screenshot of the first two sections of the report:

Please note the second arrow pointing at "Total lines processed in trace file". OraSRP is reporting 2,150,340 lines. That did not look right to me. So I went back to the raw trace file and did this:

jforonda@u20$ wc -l 4071.trc
2156850 4071.trc
jforonda@u20$

jforonda@u20$ bc
2156850-2150340
6510
jforonda@u20$

Hmm... I know that this specific trace file is 'clean' and it does not have things like those "dump from memory" lines. So it seems like there are 6,510 lines that are somehow not accounted for. What does this mean? Is it a big deal? I'll let the reader draw their own conclusions.

Next, I looked at the total elapsed time for the session. In the image above, that information would be where I drew an arrow at the very bottom of the page. For the total elapsed time, OraSRP reports 14,187.9512 seconds. Again, that didn't look right to me. So again, I went back to the raw trace file. To get a quick approximation of the session start timestamp, I did this:

jforonda@u20$ grep tim 4071.trc | head -1
PARSING IN CURSOR #5 len=69 dep=1 uid=413 oct=42 lid=413 tim=3928564402960 hv=2004533713 ad='ab3a9520'
jforonda@u20$

And to get a quick approximation of the session end timestamp:

jforonda@u20$ grep tim 4071.trc | tail -1
EXEC #15:c=0,e=1077,p=0,cr=3,cu=1,mis=0,r=1,dep=0,
og=4,tim=3942736181428
jforonda@u20$

I used the term "quick approximation" because I understand that those values may not be the absolute start and end timestamps. However, for the purpose of this blog entry, they should be good enough. I think it is safe to say the following:

the actual start timestamp can not be greater than 3928564402960
the actual end timestamp can not be less than 3942736181428

So I went back to my trusty old command line calculator. :) Note that I set the scale to 6 and divided by 1,000,000 because this is an Oracle9i trace file and so the tim value is expressed in microseconds.

jforonda@u20$ bc
scale=6
(3942736181428-3928564402960)/1000000
14171.778468
jforonda@u20$

So according to the raw trace file, the elapsed time is approximately 14,171 seconds. In fact, I can probably conclude at this point that the elapsed time can not be less than 14,171 seconds. OraSRP reports 14,187 seconds or around 16 seconds more. Okay, OraSRP's elapsed time is not less than my rough estimate but a 16-second difference does not look quite right. Hmm...

Then I noticed that OraSRP has an option to recognize idle 'SQL*Net message from client' events which was turned on in my initial run. So I made one more run with this option turned off. Here's a screenshot:

Now, the title of this blog entry partly reads "a very short initial impression" and it is getting too long :). So... can you spot the difference between the two screenshots?

OraSRP v3.0b posts to date

OraSRP v3.0b - a very short initial impression
OraSRP author replies
OraSRP v3.0b - questionable session elapsed time

Faulty FC port meets ZFS

2007-01-31T16:38:00.000-05:00

Right now, I'm in a project that is taking a lot of my time and as such, I am behind my normal readings. Whenever I'm in this situation, my mailing list subscriptions are usually the first ones to suffer.

Anyway, I just saw this post from the zfs-discuss mailing list at opensolaris.org. It was posted about two months ago but I think that it is worth posting here as well specially since a lot of my ZFS posts have been getting some encouraging levels of page hits.

The post talks about how ZFS caught a situation where different layers of the I/O stack were trusting each other to a point where data was being silently corrupted. In fact, even the administrators trusted their hardware so much. The post says (all emphasis mine):

there was no mirroring or protection at the server level, we had delegated that function to the DMX3500

I've seen this kind of "trust" in some sites. But that turned out to be a bad decision:

As it turns out our trusted SAN was silently corrupting data due to a bad/flaky FC port in the switch. DMX3500 faithfully wrote the bad data and returned normal ACKs back to the server, thus all our servers reported no storage problems.
...
ZFS was the first one to pick up on the silent corruption

Ah, silent data corruption, the kind of problem that used to keep me up at night. :)

What a pleasant surprise (the power of google)

2007-01-30T19:35:00.000-05:00

I use Google Analytics and FeedBurner SiteStats to monitor the traffic and subscriptions of this blog. This site has a very modest traffic but for the past few months now, I noticed that more and more readers are landing to this site from a google search. I also noticed that most of the searches were for silent data corruption, a topic which I had a previous entry. Currently, this specific entry is number two on google search result when one uses the search term silent data corruption.

When I tried this search today, I got this:

No wonder. What a pleasant surprise! :)

ZFS Compression

2007-01-14T00:00:00.003-05:00

This post is a result of a thread in oracle-l. It is provided to show the basics of ZFS compression.

This was done on my single core Sun Ultra 20 workstation running Solaris. The output has been slightly reformatted to make it fit the screen.

Note that this is a very simple case on a machine that is not doing any significant amount of work. Please don't use the information here as the basis of any decision you will make, specially for critical systems. Your mileage will most likely vary.

Here's a ZFS pool named p1 whose compression bit is turned on.


  root@u20# zfs get compression p1
  NAME             PROPERTY       VALUE            SOURCE
  p1               compression    on               local
  root@u20#

Let's create two filesystems under this pool and name them u (for uncompressed) and c (for compressed).


  root@u20# zfs create p1/u
  root@u20# zfs create p1/c

By default, their compression bits are turned on because they inherit this characteristic from their parent (p1).


  root@u20# zfs get compression p1/u p1/c
  NAME             PROPERTY       VALUE            SOURCE
  p1/c             compression    on               inherited from p1
  p1/u             compression    on               inherited from p1
  root@u20#

Let's turn off the compression bit for the u (uncompressed) filesystem.


  root@u20# zfs set compression=off p1/u

Let's verify that it was actually turned off.


  root@u20# zfs get compression p1/u p1/c
  NAME             PROPERTY       VALUE            SOURCE
  p1/c             compression    on               inherited from p1
  p1/u             compression    off              local
  root@u20#

Let's go get some reasonably-sized test file.


  root@u20# pwd
  /10046/tmp
  root@u20# ls -l
  total 1589543
  -r--r--r-- 1 jforonda staff 2962342454 Sep 11 2004 test_ora_17255_t.trc
  root@u20#

There... that file is an Oracle extended SQL trace file that is around 2.8G in size.

Let's make two copies of this file. First, to the uncompressed filesystem:


  root@u20# ptime cp /10046/tmp/test_ora_17255_t.trc /p1/u
  real 1:20.604
  user 0.002
  sys 14.613
  root@u20#

Then, to the compressed filesystem:


  root@u20# ptime cp /10046/tmp/test_ora_17255_t.trc /p1/c
  real 1:20.228
  user 0.002
  sys 12.637
  root@u20#

Before going further, please note the output of ptime for each of the times we copied the files. There's less than one second difference between the two out of an approximate duration of 80 seconds for each copy operation. Also, note that the copy operation to the compressed filesystem was actually faster. Hmm... interesting.

Doing an ls on the two files will show that they have exactly the same number of bytes.


  root@u20# ls -l /p1/[uc]/*
  -r--r--r-- 1 root root 2962342454 Jan 13 14:23 /p1/c/test_ora_17255_t.trc
  -r--r--r-- 1 root root 2962342454 Jan 13 14:21 /p1/u/test_ora_17255_t.trc
  root@u20#

And in fact, the Solaris digest command tells us that they have exactly the same contents.


  root@u20# digest -a md5 -v /p1/[uc]/*
  md5 (/p1/c/test_ora_17255_t.trc) = 97f86fcfdfc3f21a68ffc1892a945e77
  md5 (/p1/u/test_ora_17255_t.trc) = 97f86fcfdfc3f21a68ffc1892a945e77
  root@u20#

But the amount of space that they occupy on disk is not the same. The file residing in the compressed filesystem is around 1/3 the size of the file that resides in the uncompressed filesystem.


  root@u20# du -sh /p1/[uc]/*
  776M /p1/c/test_ora_17255_t.trc
  2.8G /p1/u/test_ora_17255_t.trc
  root@u20#

A better way to see the compression ratio is to use the zfs get command:


  root@u20# zfs get compressratio p1/u p1/c
  NAME             PROPERTY       VALUE            SOURCE
  p1/c             compressratio  3.64x            -
  p1/u             compressratio  1.00x            -
  root@u20#

Applications don't have to know that a file is compressed -- ZFS does the compression and decompression on the fly. Applications should be able to read the files normally like this:


  root@u20# tail -5 /p1/u/test_ora_17255_t.trc
  END OF STMT
  PARSE #21:c=0,e=123,p=0,cr=0,cu=0,mis=0,r=0,dep=1,og=4,tim=740525621053
  BINDS #21:
  EXEC #21:c=0,e=259,p=0,cr=0,cu=0,mis=0,r=0,dep=1,og=4,tim=740525621393
  EXEC #3:c=0,e=3416,p=0,cr=1,cu=3,mis=0,r=1,dep=0,og=4,tim=740525622047
  root@u20#
 
  root@u20# tail -5 /p1/c/test_ora_17255_t.trc
  END OF STMT
  PARSE #21:c=0,e=123,p=0,cr=0,cu=0,mis=0,r=0,dep=1,og=4,tim=740525621053
  BINDS #21:
  EXEC #21:c=0,e=259,p=0,cr=0,cu=0,mis=0,r=0,dep=1,og=4,tim=740525621393
  EXEC #3:c=0,e=3416,p=0,cr=1,cu=3,mis=0,r=1,dep=0,og=4,tim=740525622047
  root@u20#

Now, how good is the default ZFS compression compared to something like gzip? Let's find out by compressing the file that is in the uncompressed filesystem.


  root@u20# ptime gzip /p1/u/test_ora_17255_t.trc
  real 1:27.206
  user 1:19.614
  sys 2.985
  root@u20#

As expected, gzip is much better.


  root@u20# ls -hl /p1/[uc]/*
  -r--r--r-- 1 root root 2.8G Jan 13 14:23 /p1/c/test_ora_17255_t.trc
  -r--r--r-- 1 root root 99M Jan 13 14:21 /p1/u/test_ora_17255_t.trc.gz
  root@u20#

  root@u20# ls -l /p1/[uc]/*
  -r--r--r-- 1 root root 2962342454 Jan 13 14:23 /p1/c/test_ora_17255_t.trc
  -r--r--r-- 1 root root 104023266 Jan 13 14:21 /p1/u/test_ora_17255_t.trc.gz
  root@u20#

  root@u20# du -sh /p1/[uc]/*
  776M /p1/c/test_ora_17255_t.trc
  99M /p1/u/test_ora_17255_t.trc.gz
  root@u20#

The above shows the following:

Size of original file: 2.8GB
Size of file compressed by ZFS: 776MB
Size of file compressed by gzip: 99MB

Why is this so? Well, programs like gzip can take their time and optimize for compression ratio in favor of elapsed time. ZFS, on the other hand, is concerned about other things beside compression ratio so it has to strike the right balance between speed and compression ratio. That said, the ZFS man page says:

compression=on | off | lzjb

Controls the compression algorithm used for this dataset. There is currently only one algorithm, "lzjb", though this may change in future releases.

The day is coming...

2007-01-13T13:18:00.000-05:00

This is in reference to my previous entry titled "Is the day coming...".

Sun's ThinGuy says it is coming soon. You can see a demo here.

And yes, it is called win4solaris.

Is the day coming...

2007-01-11T17:20:00.000-05:00

... when you will be able to run Windows on top of Solaris x64, just like you can run Windows on top of Linux using VMWare today? As of today, Solaris is supported only as a guest operating system under VMWare.

If Solaris can be a host operating system to another OS, then it is possible to use some of the Solaris features on the hosted OS. For example, today, it is possible to use DTrace on a Linux system running on a Solaris BrandZ Zone. In a similar fashion, if one can run Windows on top of Solaris, it is then possible for Windows to indirectly take advantage of features such as DTrace and ZFS. And with things like resource management, it then becomes possible to cap the resources used by that instance of the Windows OS. The thought of being able to do this excites me no end. :)

Anyway, I've always wondered what Sun is going to do in this area. A few years ago, we had a product called SunPCi. This product is a combination of hardware and software. The hardware component was a card that goes into a SPARC box. This card had an AMD processor. With this hardware/software combination, one can then install Windows and some Linux distros on top of Solaris. It was similar to VMWare in so many ways. I used SunPCi a lot and I loved it.

When Solaris x86/64 was announced, I thought that Sun would then take the SunPCi software component and make it run on Solaris x64/86. But that has not happened.

Then, last October, ThinGuy mentioned that Sun is working with Win4Lin. Today, I see this.

Hmmm.... win4solaris, anyone?

Merry Christmas/Season's Greetings - in 38+ dialects/languages

2006-12-20T09:55:00.000-05:00

Five years ago, somebody at Sun started an email thread about how the different forms of this season's greetings are said in different languages. This is what we came up with.

If you want to add, correct, or translate something, please leave it as a comment and I will update the list.

Updates:

Added translation to the French greeting (courtesy of my daughter).
Corrected/added/translated German (courtesy of Alex Gorbachev).
This post was accepted to Darren Rowse's ProBlogger Group Writing Project. This, in turn, generated several links to this post, and I'm happy to see that it made it to someone's favorite picks.
Added/corrected Russian (courtesy of Alex Gorbachev).
Added Oracle (courtesy of Alex Gorbachev).
I'm submitting this post to Darren Rowse's ProBlogger Group Writing Project.
Changed all references to 2002 to 2007. I purposely didn't do this initially because I was not sure if some of the non-translated greetings mentioned the year in words, as in: two thousand and two instead of 2002. If somebody wrote "two thousand two" in Mandarin, I wouldn't catch it. :)
Added Russian and Latvian (courtesy of Edgar).

So with that, here's a Merry Christmas / Happy Holidays / Season's Greetings to you!

Chinese
koan hay fat choi (wishing you prosperity in the new year)
French
à tous et toutes, UN TRES JOYEUX NOEL ET UNE BONNE ANNEE 2007! (To all, A VERY MERRY CHRISTMAS AND A HAPPY NEW YEAR 2007!)
English
Merry Christmas and a prosperous 2007
Italian
Buon anno, Happy New Year
Tagalog (Philippines - national language)
Maligayang Pasko at Manigong Bagong Taon (Merry Christmas and a Prosperous New Year)
Ilocano (a dialect in northern Philippines)
Naimbag a Paskua ken Naragsak a Baro a Tawen (Merry Christmas and a Happy New Year)
Dutch
Iedereen een gelukkig kerstfeest en een goed oud en nieuwjaar toegewenst.
- Wishing all of you a Merry Christmas and a happy Old and New Year.
Persian 1
etoon mobarak bashe, baba in che va'zesheh to ham maro film kardi
Spanish
Feliz Navidad
Bulgarian
Vessela Koleda & Chestita Nova Godina 2007
Russian
S Rozhdestvom (Merry Christmas)
S Novym 2007 Godom ( Happy new year 2007)
S Novym Dve Tysyachi Sed'mym Godom. ("d'" means soft d)
Latvian
Priecigus ziemassvetkus (Merry Christmas)
Laimigu jauno 2007. gadu (Happy new year 2007)
Bengali
'Shubho Naboborsho' (Happy New Year in Bangla)
Portuguese
Feliz Natal e Próspero Ano Nôvo!
Urdu
Aap Sabko Naya Saal Mubarak ho
Persian 2
Azizan; eideh shoma mobark- dombeh shomah secharak.
Hebrew
Hag sameah (Happy Holiday)
Shana tova ve-metuka (sweet and happy year)
Indonesian
Selamat hari natal dan tahun baru 2007!
Lebanese
Milad Majid wa Aam Saeid! (Merry Christmas and Happy New Year)
Yiddish
A Git Yoor - (Happy New Year in Yiddish (Oy, we don't have Christmas))
Hindi
"Naye Saal kee Hardik Shubhkamnayen" (Heartfelt Best wishes to all for the New year)
Danish
Glædelig jul og godt nytår (Merry Christmas and Happy New Year!)
Albanian
Gezuar Krishtlindjet! Gezuar Vitin e Ri 2007!
Tamil Version 2.62
pudhu aandu valthukkal 2007
Marathi
Navin Warsha sarwanna Sukhache wa Samruddhiche jaavo (Wish you all Happy and Prosperous New Year)
German
Schöne Weinachten und Gutes Neues Jahr! (Literal: "Nice Christmas and Good New Year" but you would say "Merry Christmas and Happy New Year.")
Ein Fröhliches Weihnachtsfest und einen glücklichen Rutsch ins Neue Jahr! ( Means something like: "Happy Christmas festival and successful journey in the new year.")
Turkish
Yeni Yiliniz kutlu olsun (have a happy new year)
Bosnian, Croatian and Serbian
SRETAN BOZIC AND NOVA GODINA (Happy Christmas and New Year)
Greek, phonetic
Kala Hristuyenna ke eftihismeno to neo etos!
Romanian
Craciun bun si an nou fericit!
Vietnamese
Chuc Mung Nam Moi (Happy New Year)
Gaelic (Irish)
Nollaig chridheil agus Bliadhna mhath ùr! or Beannachtaí an tSéasúir (Greetings of the Season)
Zulu
Sinifesela Ukhisimusi Omuhle Nonyaka Omusha Onempumelelo (Merry Christmas and a Happy New Year)
Slovak
Vesle Vainoce a Stastlivi Novy Rok! (Merry Christmas and Happy New Year!)
Kokani (India)
Tumka saglyak natalachi aani aanandache nave varsh zav.
Swahili
Kuwa na Krismasi njema (Merry Christmas)
Lugandi (Uganda)
Amazalibwa Agesanyu or Nkwagaliza Sekukulu Omulunji
Ukranian
Z Novym Rokom ta Rizdtvom Hrystovym!
Ukranian 2
Chrystos Rodihvsya ee Shchaslihvoho Novoho Roku (Christ is Born - have a Happy and Prosperous New Year)

Some folks added these funny ones:

Binary
0100100001100001011100000111
C
printf("%s %s %s.\n", "Jesus", "is the Reason", "For the Season");
Oracle
SELECT 'Happy New Year ' || TO_CHAR(ADD_MONTHS(SYSDATE,12),'YYYY') FROM DUAL;
Klingon ;-)
Cho chey kr'uug my-uk pogn'r
The Dyslexic Canadian
!samtsirhC yrreM
Pig Latin
Appyhay Ewnay Earyay
Elven
valin winya atendea (Tribute to Lord of the Rings)

T2000 - eBay's Database Platform of Choice for 2007

2006-12-14T10:09:00.000-05:00

I've always loved it when a quote comes directly from a customer. Here's a recent one that describes the T2000 as scary fast :)

“ Niagaras run much cooler than our previous servers, plus they are scary fast. They’ve cut our total cost of operations [TCO] through lower acquisition and management costs. The Sun Fire T2000 will be our database platform of choice in 2007. ”

— Heather Peck, Infrastructure Manager, eBay Inc.

eBay mainly uses Oracle.

You can read about it here.

Why stop there?

2006-12-09T07:48:00.000-05:00

Today, this article from The Register caught my attention:

US politicos demand cool servers

US politicians have joined the great call for energy-friendly data centers, as the Senate this week approved a bill that promotes low-power server technology.
Senators unanimously pushed through the legislation that asks the Environmental Protection Agency (EPA) to study the energy consumption of data centers. The EPA has been instructed to bone up on items such as low-power chips, power supplies and energy distribution. It will also consider how incentives could be used to promote power-efficient data centers in the government and private sectors.

Energy efficient computing is becoming more and more of a topic these days. I'm glad to see that politicians are finally putting in place legislation to promote green/cool computing.

Even utility companies have started to offer incentives to promote green/cool computing. A few months ago, PG&E teamed up with Sun to offer up to US$1,000 rebate for servers:

Sun Teams with PG&E On First Ever Rebate For Servers, Up To $1000 In Savings

Sun Microsystems (NASDAQ: SUNW) today announced that Pacific Gas and Electric Company (PG&E), one of the largest natural gas and electric utilities in the United States serving 350,000 California businesses, is offering significant rebates as part of a new energy incentive program for computer servers. The CoolThreads technology Sun Fire T1000 and T2000 servers -- which are three to five times more energy efficient than competing systems and take up less space(1)-- were the only servers to qualify for the rebate program. As part of PG&E’s Non Residential Retrofit program, customers replacing existing equipment with these eco-responsible servers can receive a cash savings between $700-$1000 per server or up to 35 percent when combined with the Sun Upgrade Advantage Program(2). This is the first-ever incentive rebate offered by a public utility company for servers.

Very nice and cool. I think its definitely a step in the right direction.

But why stop there? Right now, the focus is on servers. But there are also millions (billions?) of power-hungry desktops that are out there. How much power does an average desktop, without the monitor, consume? 200 watts? More? At Sun, our standard "desktop" hardware consumes as low as 4 watts. Yes, as low as four watts.

We used to call the Sun Ray an ultra-thin client to differentiate it from other thin client solutions. Other terms such as Display Over IP (DOIP) were also used to describe the technology but some of us sometimes replaced the meaning of D to Desktop so that DOIP would mean Desktop Over IP. Because that's what it really is -- our entire desktops are being delivered to us over the network. Then, just recently, "ultra-thin client" has been replaced with Virtual Display Client.

I'll be blogging more about this soon. In the meantime, you can look it up here.

Scaling out?

2006-12-08T18:40:00.000-05:00

Kevin Closson has a couple of interesting posts about Sun gear and Oracle here and here.

In the second post, he says:

But even for the “heavy” databases, it wont be long until they too only require some of the cores that a single socket offers.
...
OK, remind me again who actually needs to “scale-out” a database to multiple servers?

Very good point. Very good question.

I personally believe that a lot of the companies who use Oracle RAC today don't need it.

As to when "heavy" databases will need only some cores that a single socket offers, I'm already seeing this situation today. While a lot of people talk about scaling out - and some companies are actually scaling out - I also see companies that are consolidating their databases on more powerful machines. Right now, I'm working on a project for one of the big telcos in North America. You would think that these guys would be scaling out like crazy. On the contrary, they are consolidating all sorts of things, including big databases, on more powerful machines.

Done with setup mode

2006-12-03T20:01:00.000-05:00

I've finally moved all my posts from my old blog to here. I could not find a blogspot feature to easily import the data from a Wordpress blog. I did not look that hard since I only had a few posts so I just did the move manually.

I am aware that each of the posts that I moved over resulted in a feed to those who subscribed via aggregators so my apologies to you if I flooded your aggregators the past couple of days. The worst should be over now. :)

Just to let you know, I’m planning to post a little bit more frequently soon. But since I do all this outside of Sun activities during my personal time, please bear with me as I try to get into a rythm amidst all the work I’m doing in my day job -- a rythm that will still allow me to keep a balance in my personal life.

As of now, I plan to blog mostly about Sun technologies and databases. I’ve also been taking a deeper look at the Oracle Extended SQL Trace file (event 10046, level 8+) and I might blog about that as well. So please stay tuned.

And lastly, please note that I've added a blog subscription by email. I am aware that a lot of people still prefer email and if you are one of them, the email subscription is for you. :)

NewsGator: Why?

2006-11-30T21:13:00.000-05:00

I was playing around with different news readers/aggregators when I remembered that I have an existing NewsGator account that I set up a while back. So I went to the NewsGator site and tried to login but I forgot my password. Like most sites, NewsGator has a "Forgot your password?" link. Using that function sent me this email:

From: support@newsgator.com

To: [My Email Address here]

Subject: NewsGator Online credentials

Date: Thu, 30 Nov 2006 19:04:09 -0700
As you requested online, we are sending your NewsGator
Online credentials to you.  They are as follows:

Username: [My Username was here]
Password: [My unencrypted password was here]

Customer Support
NewsGator Technologies

Huh?

If this was a password reset or new account creation, that would have been acceptable -- as long as I am then required to change it the next time I log on. But what they sent me is a password that I created months ago.

This could only mean that in NewsGator's servers, my password is stored either in unencrypted format, or if it is encrypted, it is reversible. I know that a lot of smaller sites do this... but NewsGator? What happened to one-way encryption?

Comments? Good, bad, tolerable, unacceptable?

Still in setup mode

2006-11-30T09:39:00.000-05:00

As of yesterday, my main blog was still at http://jforonda.vintarinian.com. However, I was trying to revive this blog account because I was planning to move back here (long story that I will explain later). I was playing around with feedburner for my feed and to my surprise this morning, there are already people subscribed! Well... thanks! :)

So to you who have subscribed, please bear with me as I move my posts over. Because of you who subscribed, I am now making this the main blog as of today.

The http://jforonda.vintarinian.com blog will be decommissioned before December 15, 2006.

TOUG folks meet Solaris 10 :)

2006-09-26T11:52:00.000-05:00

I am very happy that even if we (Sun) were not able to make a formal presentation at the Toronto Oracle Users group yesterday, there were a lot of people who stopped by our booth. The number of people, the kinds of questions asked, the comments given, and the lenghts of conversations we had -- all of these gives me a very good feeling about how Sun and Solaris are being viewed these days.

So... as promised, here are the initial notes and links to the major points we talked about yesterday, together with some brief comments:

Solaris is Open source. FAQ for the CDDL license is here.
While it is true that Solaris traditionally ran on SPARC, it also now runs on x86 (32-bit) and x64 (64-bit) platforms. The SPARC and x86/64 versions are built from a single source tree.
Sun sells SPARC as well as Opteron based hardware and of course, these are all certified against Solaris. However, it is not correct to think that you have to get the hardware from Sun in order to run Solaris. Solaris x86/64 runs on a lot of existing hardware; I've been running it on my Toshiba Tecra M2 laptop for more than a year without any problems. Here's the Hardware Compatibility List. Another way to see if your hardware can run Solaris 10 is to test it using the Solaris x86 Installation Check Tool.
The DVDs that we gave away yesterday contain Solaris 10 Update 1 (S10U1). S10U1 includes Zones but not ZFS. The latest supported release of Solaris 10 is Update 2 (S10U2) and it includes ZFS. You can download S10U2 here. Solaris license is free. Support pricing starts at $120/yr.
Non-RAC Oracle database is certified on Solaris Zones. There's a note on Metalink about this (I'll put the link later). RAC is not certified with Zones but RAC is certified to run from the global Zone.
Solaris Zones home page is here and the Zones FAQ is here. The Apache Software Foundation uses Solaris Zones.
Solaris ZFS home page is here. I also have a couple of ZFS posts here and here.
DTrace home page is here. DTrace has been ported to Mac OS X and there are talks that Apple is also considering ZFS for OS X.
Oracle's preferred OS for 64-bit development and deployment environment is Solaris. Here's the press release from Oracle's web site.

I think that's it for high-level notes and links. Now, quite a number of you also asked if I could post some tutorial type of examples on what we talked about yesterday. I will be doing that as soon as my schedule permits. So please check this blog often. Better yet, please subscribe.

That was a great day yesterday. Thanks for dropping by and I hope that at the very least, you now see why a lot of people say that Solaris 10 rocks! :)

Sept 25 - Free format Solaris/Oracle demo at Toronto Oracle Users Group

2006-09-20T12:02:00.000-05:00

I was asked to man Sun's booth at the Toronto Oracle Users Group's Oracle Day 2006 this coming Monday, Sept 25. If you happen to be attending this event, drop by the Sun booth and see how you can use Solaris 10 features with Oracle.

I don't have any canned presentation. This will be free-format -- just come by and ask your questions and we will answer what we can.

See you there.

ZFS - Another way to recover from a screw up... :)

2006-08-31T22:29:00.000-05:00

Doug Burns wrote in his blog how he accidentally messed up some of his Oracle database development environments. Doug, I remember that a little while back, we had a short chat over at Christo Kutrovsky's follow-up blog entry to my ZFS post.

If you've had the chance to install Solaris 10 update 2 (6/06 release), it would be a good time to try ZFS, specially on development environments. ZFS has a snapshot feature and this is one of the ZFS features that I really love.

For example, I use this to create snapshots of entire filesystems. In your case, if there was a snapshot of /ora/data2, recovering the data from the snapshot would have been very f-a-s-t (most likely seconds). I routinely do this with my development environments.

I also use ZFS snapshots in place of the OS-level copy command when doing Oracle hot backups. When done this way, hot backups can be completed in seconds no matter how big the datafiles are.

Another place where I use ZFS snapshots is when I upgrade Oracle or any other software that is mounted under ZFS. If something does not work as I expect, a zfs rollback will quickly (again, seconds) bring order to a messy situation. :)

But my favorite use of ZFS snapshots is to snapshot entire Solaris 10 zones. Sometimes, these zones host Oracle databases. If something gets messed up, I simply issue a zfs rollback command and the entire zone goes back to its state when the snapshot was taken. Very fast and very neat. :)

I've been meaning to make a more detailed post about this for a while now, complete with examples for Oracle. Maybe now is a good time for me to do that. I hope I can make time to write something soon.

2 suggestions for Boston

2006-08-16T18:10:00.000-05:00

My family and I were just in Boston for the first time as part of our summer vacation.

Boston is a beatiful city. Of course, different people have different tastes, but to me, Boston is easily one of the most beatiful American cities I've seen. Great architecture, people, and of course, food!

I love to walk and I take public trasportation when I can because doing so gives me the opportunity to have a closer interaction with the people. 'Up close and personal' is what others would say. :) So I was happy when I learned that Boston has a good subway system and that it is a city that encourages walking. They have walking tours and they also have 'Walking Boston' maps and guides.

We all had a great time but taking Boston's subway system and walking its city tours made me think that even great cities like Boston can make small changes that can make the experience even better. So here are my two small suggestions for Boston:

1. Please stop selling those automated subway tickets specially to tourists. Why? Because not all the subway stations have machines to process those tickets. Before we bought our subway tickets, I spoke with a subway employee and he advised us to get the automated tickets. I asked if that kind of ticket would be honored in all stations and he said yes. But he was wrong; there are several stations that don't have machines to process those tickets. We were forced to buy additional tokens at one station even if our tickets were not yet completely used up because the attendant simply refused to let us through. He said something like: "We are aware that it is a problem but we were told that we are not supposed to let people with automated tickets through". Huh?? Fortunately, attendants at other stations were willing to "bend the rules" and they let us through.

So, Boston... at the very least, please put big signs on automated ticket dispensers saying that the automated tickets won't work everywhere. And if there are readers who are planning to take the subway to go around Boston, I think it is a lot better to just buy the subway tokens instead of the automated tickets.

2. I'm not sure how much Boston embraces recycling but if they do embrace it, I think they should put more recycling bins specially along the walking trails and subway stations. Majority of people who love to walk will be carrying some form of drink with them. I held on to my empty water bottle for around one hour hoping that I will see a recycling bin. I did not see one and when I finally made the decision to toss it to a garbage bin, I noticed that the garbage bin had a lot of empty plastic bottles in it. Now, I'm not saying that there were no recycling bins at all. I'm just saying that if there were, I did not see them despite the fact that I was consciously looking for them.

That said, I have to say again that Boston is a very beatiful city. There are other parts of the city that I want to see that's why I'm looking forward to my business trip there in a couple of weeks.

A tough day @ Sun

2006-08-04T00:43:00.000-05:00

Today was a very tough day at Sun. Once more, I see some friends affected by the announced reduction in force of up to 5,000 employees worldwide.

To my colleagues who have lost their jobs, hats off to you. Please take care of yourselves and keep in touch.

It's not clear if it is all over. Anything can still happen.

And with this, I unplug myself from this blog for a couple of weeks of vacation.

If my blog looked funny over the weekend...

2006-07-31T14:13:00.000-05:00

... it was because I messed up something while playing around with some WordPress themes. It looked fine from my computer but then a reader using MS Internet Explorer informed me that my pages didn't look like they used to.

My main workstation is running Solaris. I guess that's what I get for not using MS Internet Explorer.

Sorry about that. I'll be more careful next time.

I hope they can make this work

2006-07-28T09:49:00.000-05:00

This Wired article just caught my attention.

Powering Up, One Step at a Time

Excerpts:

British engineers are converting street vibrations into electricity and predict a working prototype by Christmas capable of powering facility lights in the busiest areas of a city.
"We can harvest between 5 to 7 watts of energy per footstep that is currently being wasted into the ground," says Claire Price, director of The Facility Architects, the British firm heading up the Pacesetters Project. "And a passing train can generate very useful energy to run signaling or to power lights."
...
Gilbert is working with hydraulic-powered heel-strike generators, which he believes could be installed in the floors of busy public places like subway stations. Those stations typically capture the footfall of 20,000 commuters an hour during peak usage -- multiplied by 5 to 7 watts a person, that's more than enough to power a building's lights for the day.
...
The prototypes are scheduled for assembly by December 2006.

I think it's a cool idea. I hope they can make it work.

India rejects 'One Laptop per Child'

2006-07-26T23:39:00.000-05:00

The Times of India has reported that India's Ministry of Human Resource Development has rejected the idea of One Laptop per Child.

Excerpt:

The HRD ministry has rejected the idea of 'one-laptop-per-child' (OLPC) being aggressively marketed by Nicholas Negroponte of MIT Media Laboratory. "India must not allow itself to be used for experimentation with children in this area," the ministry has said.
The ministry's detailed objection based on technical, social and financial grounds was sent to the Planning Commission two weeks ago.

It is not clear what the technical objections are; there was nothing mentioned in the article. As for me, I have my own personal reservations about OLPC's approach. At this time, I am not convinced that they are headed in the right direction. Why? Here's something from OLPC's FAQ page:

What is the $100 Laptop, really?
The proposed $100 machine will be a Linux-based, with a dual-mode display—both a full-color, transmissive DVD mode, and a second display option that is black and white reflective and sunlight-readable at 3× the resolution. The laptop will have a 500MHz processor and 128MB of DRAM, with 500MB of Flash memory; it will not have a hard disk, but it will have four USB ports. The laptops will have wireless broadband that, among other things, allows them to work as a mesh network; each laptop will be able to talk to its nearest neighbors, creating an ad hoc, local area network. The laptops will use innovative power (including wind-up) and will be able to do most everything except store huge amounts of data.

Here's my concern: considering that the target users are young children (6-12 year olds, if I'm not mistaken), how wise is it to make a machine that is run by a full-blown operating system? Do we expect these kids to "play system administrators" on their laptops? Do we expect them to know how to patch their systems when there are security vulnerabilities in the operating system or application programs? Or are they planning to have professional system administrators push the updates to these laptops? Is that really going to succeed?

I am not picking on Linux just because I am a Sun employee. I will still have the same reservations even if they use Solaris as the operating system. The fact is all operating systems will have their own vulnerabilities even if these operating systems are stripped down of unnecessary services. And when these vulnerabilities are discovered, they have to be dealt with. Besides, no matter what operating system is used, there are still application programs that will also have their own vulnerabilities.

OLPC is talking about deploying millions of these laptops. Wouldn't it be better if they are engineered in such a way that the end users will no longer have to worry about securing their individual laptops?

I think it makes sense to study the viability of a thin client model -- a device that has no operating system to begin with. That way, there is nothing to maintain on the laptop itself.

Things that caught my attention: biometric spoofing

2006-07-21T11:52:00.000-05:00

Yesterday, I started a category that I called Things I take for granted. Today, I am starting another category. This time, I will call it Things that caught my attention.

First to make it to this category is biometric spoofing. I've had this lingering concern about biometrics for a while now. If people start using biometrics as a means to authenticate, what happens when someone gets a hold of the digital representation of your fingerprint or your iris? A password is easy to change but how easy is it to change your fingerprint? How about your iris?

Today, my news aggregator caught this article from Slashdot . The article pointers lead to ZDNet Asia:

Crime of the future--biometric spoofing?

Excerpts:

Watch where you leave your fingerprints--soon they could be the target of thieves looking to break into your bank account.
Although biometric security systems--using fingerprints, iris scans and facial recognition--are only just now entering the mainstream, they are likely to be common within a few years.
And as soon as biometrics begin to be used to protect bank accounts or benefit systems, crooks will start looking at ways of breaking into them
....
"We are leaving our prints everywhere so the chance of someone lifting them and copying them is real.
"Currently it's only researchers that are doing spoofing and copying. It's not a mainstream activity--but it will be. It's just human nature; if it can be done it will be done if you can achieve some benefit from it."
Different biometrics may be attacked in different ways. For example, researchers have proved in the past it is possible to trick fingerprint systems with fake fingers made of gelatine.
Similarly, would-be thieves could try to spoof facial recognition systems with photos, videos or facial disguises in order to get access to the systems or information they protect.
Part of the problem is that many of the biometrics used by these systems are easily visible.
Toth warned: "Many people are trying to regard biometrics as secret but they aren't. Our faces and irises are visible and our voices are being recorded. Fingerprints and DNA are left everywhere we go and it's been proved that these are real threats."
In response vendors are building tighter security into their biometric systems--for example to check that a finger has a pulse, or that a real iris is being presented rather than a photo.

For now, I don't trust biometrics as a means to authenticate. Maybe that's because I don't fully understand the technology behind it. But I like it better when you authenticate with something you know, something stored in your brain -- like a password that will in turn generate a random password. That way, the master password can easily be changed when there is a security breach. The random password, on the other hand is just that -- a randomly generated password that is valid only for a few seconds. This is the type of authentication that we use at Sun. Maybe one day, I will post something about that.

Perhaps, one day, biometrics will become really secure. But I will wait and see. Until I'm convinced, I won't sign up for any service that uses biometric authentication.

Things I take for granted: syndication

2006-07-20T16:59:00.000-05:00

Today, I'm starting a new category that I will call Things I take for granted. This category will contain posts about things that are almost second nature to me and yet may be completely foreign to other people. Of course, this is never meant to put anyone down. I'm sure that things that doctors take for granted will be foreign to me as well.

First to make it to this category is syndication. Most of my peers in the IT industry will take syndication for granted but to a lot of people, it is a complete mystery. Just yesterday, somebody asked me what those syndication buttons are for. Quite a few people have also asked me where I get the time to read all sorts of things. The secret: syndication and aggregators.

For those who are new to syndication, here's a good introductory article:

What is RSS/XML/Atom/Syndication?

Excerpts:

Syndication is the process of using RSS/Atom for automated updates, another way of getting the information you want. You no doubt have a list of web sites you browse daily for updates, whether they’re stored in your bookmarks or your head. If you find yourself loading 20 or 30 sites a day, and you notice if a few stop updating as frequently, you’ll inevitably stop checking them.
What if there were instead some way to have your list of bookmarks notify you when the sites you read have been updated? You wouldn’t waste time checking those that haven’t. Instead of loading 30 sites a day, you might only need to load 13. Cutting your time in half would enable you to start monitoring more sites, so for the same amount of time you originally invested in checking each site manually, you may just end up end up following twice as many.
Syndication provides the tools to do this. A news reader, or aggregator as they’re also known, is a program or a web site that automatically checks your list of bookmarks (which you only have to set up once) and lets you know what’s new on each site in your list.
.....

As an analogy, the news reader acts like a customizable newspaper. You can pull a variety of content from a growing number of sources into one place, to be read however you choose.
.....
The only stipulation is that the source must provide a feed....
If this introduction has whet your appetite, the next step is to grab a news reader and start playing.
.....

So there you go. If you have a Yahoo, Google, MSN, or AOL account, you can instantly take advantage of syndication. If you don't, there are free online aggregators like Bloglines, Rojo, NewsGator, and Netvibes. Also, some browsers like Firefox have builtin support for syndication.

Go ahead, try it so next time I post something, you will be notified right away. :)

Blog moved: Blogger.com -> WordPress

2006-07-18T20:10:00.000-05:00

I've decided to spend more time blogging and since I am not very happy with blogger.com, I thought that I might as well get out of there while my blog is just starting. I feel that blogger.com is too restrictive. I have also read some discouraging posts here, here, here, and here.

So before I become one of blogger.com's problem statistic, I did some research and I decided to go with WordPress. Since I have a web host that is just sitting around and I enjoy doing installation and configuration anyway, it made sense that I host it myself. I feel that hosting it myself will give me more control.

Installation of WordPress was smooth. It even has an import function so I was able to import my entire blog. However, I noticed that the import was not perfect. Some of the blockquotes and links were not properly moved over. But I only have a few entries so far so I can live with that. Those should be easy to fix.

The thing that took the most time, though, is fiddling around with themes and plugins. After a few hours of fun, I am now at a point where I'm happy with the look and feel of the site.

There are still some things to do like redirecting my blogger.com pointer and setting up automated backups of this blog. I also need to put in the syndication stuff.

But it's 8pm so this should be good enough for now. :)

What's wrong with this picture?

2006-06-16T11:35:00.000-05:00

Every now and then, I get pulled into computer performance improvement projects. In a lot of places I've been to, some people try to solve performance issues by throwing hardware at the problem. Most of the time, this does not solve their original problem. Sometimes, it even makes things worse.

What I usually see is that computer systems are doing too many unnecessary things. They are bogged down by doing things that they are not supposed to be doing in the first place. When I get into this situation, I focus on unnecessary workload reduction. In most cases, reducing or removing unnecessary workload solves the problem.

So when I saw this, my first reaction was "Why?".

At Sun, we take a different approach. We use energy efficient servers and desktops. The result is that not only do we save power to run our computers, we also don't need as much power to cool the space.

Unnecessary power consumption reduction.

Clearing some misconceptions about Sun and Solaris

2006-06-07T05:32:00.000-05:00

I've been following some threads on the Oracle-L list recently and I noticed that there are some misconceptions about Sun and Solaris by some of the members of the Oracle-L community. I made some posts to set the record straight. Let me just summarize them here:

1. In a thread titled "Anyone used 10g Release 2 (10.2.0.1.0) for Solaris Operating System (x86-64)", a member of the list asked:

I have been told I may have to install and support Oracle Database 10g Release 2 (10.2.0.1.0) for Solaris Operating System (x86-64). I have plenty of experience with Solaris for Sparc but the database for the x86 version (64 bit) was just released in March I believe. I am curious about the experiences of anyone that may have this configuration. Any information would be greatly appreciated.

and one of the replies was:

You're paying the cost of using exotic platforms. I sincerely doubt that anybody will be able to respond to your question.

This surprised me because I know that late last year, Oracle choose the Solaris 10 Operating System as its preferred open source 64-bit development and deployment environment. So I replied and provided a link to the press release. I also pointed out that Solaris is open source because I get the impression that to a lot of people, "open source = Linux" and that Linux is the only Unix-like OS than can run on x86/64 hardware.

The original question also implies that some people think that Solaris 10 SPARC and Solaris 10 x86/64 are entirely different products. The fact is that both are built from the same source tree and so experience with Solaris SPARC will carry over nicely to the x86/64 platform. I know this for a fact because I use both Solaris SPARC and Solaris x86/64.

2. In another thread with the subject "Oracle's relationships with expert DBAs (and the rest of us mere mortals)", another member of Oracle-L said:

Also, I'd suggest you replace those 20 Sun servers with a couple of Opteron-loaded standard boxes. Just for your main production environment, mind you. Might be easier to handle a couple of standard PC boxes than all those refrigerator-size monsters out in the big, cold room.

I understand the reference to "refrigerator-sized monsters" for it is true that Sun has some of these high-end machines and I have used some of these machines in the past. So I pointed out that maybe, the person had this kind of machine in mind. But if, for whatever reason, one prefers Opteron based machines or smaller, less expensive, and cooler running machines in general, then Sun has its Galaxy (from $745) and CoolThreads (from $2,995) line of servers.

My experience is not unique. Sun's new CEO, Jonathan Schwartz, recently said in his blog:

I was with a big potential customer yesterday - in the Fortune 100. After a day of briefings from our technical folks, I joined the meeting to see how we were doing. I asked him and his team how much of what they'd seen was new to them.
He said, "about 70% was a complete surprise."
Ouch. That's not good.
To test, I asked, "before today, did you know that Solaris was open source, or ran on Dell, HP and IBM hardware, not just Sun's?" "Nope."
And like I said, this was a Fortune 100 opportunity.

Silent data corruption

2006-06-04T23:17:00.000-05:00

Below is a slightly modified version of a post that I made to my high school batch mailing list. We were having a discussion about OS vulnerabilities and the importance of making backups.

--------------------------------------

Years ago, I was a full time database administrator. If there is an unforgivable sin in database administration, it is losing data. Database performance may be slow... that's okay, that's fixable. But if I lose data, I should probably be making sure that my resume is up to date. :)

And so I tried to learn everything I can about database backup and data protection. Full backup, incremental backup, cold backup, hot backup, replication, standby databases, mirroring disks, breaking the mirrors, hardware raid, software raid, volume managers, hot spares, etc. I made sure I studied and practiced everything until I was satisfied that I knew what I needed to know. But not only that, I also made sure that I knew how to recover in case of disaster. I would occasionally take some of my backup and try to do a restore just to make sure that I can really recover should there be a need to do so. To put things in perspective, I was dealing with a database that was approaching 2 terabytes in size so this was not a case of simple backup/recovery scenario.

But as one of us pointed out, disks fail. If one of my disks should fail, my hope was that it would fail in an unmistakable way. I hoped that if a disk failed, the other parts of the system would know in an unmistakable way (like an I/O failure) such that the safeguards that we put in place (like hot spares) would kick in. If this kind of failure happened, we can deal with it and we would know how to fix it.

But what kept me up at night was the thought that there is a possibility that at one point in time, my data has experienced what we call silent corruption. This can happen with hard disks. But it does not stop there. Silent corruption can also happen when you have a hard disk controller failure. I can also imagine that there could be bugs with device drivers. I also know that some disks have firmware and that, too, can be a source of problems. Again, if these errors result in a clear I/O error, then that's the best thing that could ever happen because then, we would know that there is a problem and we could deal with it. But what if it is an undetected error? And what if it goes undetected for a long period of time?

So if my database had a silent corruption, how do I know when it happened? How do I know which of my backups is good? A silent corruption could mean that all my backups are useless because it is possible that the problem is older than my oldest backup. Sure, there are database utilities to verify a database for inconsistencies but when does one run these utilities against a 2TB database? Before I do a backup? How long will that take? Will it fit into the backup window?

Anyway, enough with databases. The problem with silent data corruption is obviously not unique to databases. It is a problem that can happen with any type of data.

So how does one safeguard against this type of problem? There could be other ways but one way is to use Solaris with ZFS. :-) ZFS has self healing features (look at slides 13-15). There's also a screencast demo here

Some web sites like strongspace are already using it even before its official release.

How stable is it considering that it is not even officially released? Read this blog from the lead test engineer:

Testing ZFS reminds me of when I was working for IBM at NASA on the Space Shuttle Flight Software Project....

And here's one more from another test engineer.

All in all, this means that we put our code through more abuse in about 20 seconds than most users will see in a lifetime. We run this test suite every night as part of our nightly testing, as well as when we're developing code in our private workspaces. And it obeys Bill's first law of test suites: the effectiveness of a test suite is directly proportional to its ease of use. We simply just type "ztest" and off it goes.

For those who want to learn more, here is a good starting point. ZFS home page is here.

Hello world

2006-06-03T21:33:00.000-05:00

My name is James Foronda. I work for Sun Microsystems' US Data Center Solutions - Enterprise Migration practice.

From:	support@newsgator.com
To:	[My Email Address here]
Subject:	NewsGator Online credentials
Date:	Thu, 30 Nov 2006 19:04:09 -0700